The effect of carvedilol on morbidity and mortality in patients with chronic
Authors: Packer M, Bristow M, Cohn J, Colucci W, Fowler M, et al.
Source: New England Journal of Medicine. 334:1349-55.
May 23, 1996.
Institutions: Multi-institutional in the US.
Financial support: SmithKline Beecham, Boehringer-Mannheim.
Multiple lines of evidence point to activation of the adrenergic nervous
system as a concomitent and aggravating factor in congestive heart failure.
Adrenergic blockade is thus a plausible approach to therapy, but the reliance
of the failing heart on sympathetic tone to maintain function leads to
a risk of dramatic decompensation if adrenergic support is abruptly withdrawn.
Thus, trials of beta-blockade in heart failure have been carried out very
slowly and with caution. Initial results have been positive but have shown
mainly symptomatic benefit rather than reduced morbidity and mortality.
This randomized, placebo-controlled trial, multi-institutional in the US,
was designed specifically to look at morbidity and mortality. The agent
investigated was carvedilol, a non-selective beta-blocker with alpha-1
blocking (vasodilating) and anti-oxidant properties.
Subjects: Patients with ischemic and non-ischemic dilated cardiomyopathies,
EF <: 35%. of at least three months duration. Exclusion criteria were
designed to eliminate recent acute cardiac events, patients with major
arrhythmias, very high or low BP, hepatic, renal or other life-shortening
illnesses and treatment with calcium blockers, beta agonists or blockers
and the class IC and III anti-arrhythmic agents.
Intervention: Patients were screened during a three-week period
and stratified according to the results of a 6-minute corridor walk test.
All patients then underwent an initial two-week open-label trial of 6.25
mg carvedilol twice daily. Those who were able to tolerate the drug went
on to randomization. The figure below shows the numbers of patients randomized
in the various groups. Patients randomized to carvedilol were titrated
up towards 50 mg bid, except for a subgroup randomized to a dose-ranging
study, where the maximum doses were as shown in the figure. Stratification
was according to the distance in meters the patient was able to walk during
Analysis: The plan was to enroll 1101 patients; the primary endpoint
was mortality, the secondary endpoint was hospitalization for CHF or other
cardiovascular disorder. Follow-up was to be for 6 months (12 months in
the mild CHF group). A monitoring committee was to terminate the study
early if a "clinically important treatment effect" (not specified ahead
of time) was observed.
Enrollment began April 29, 1993; the trial was halted on February 3, 1995
after randomization of 1094 patients because of a significant effect of
carvedilol on survival.
1197 patients entered the two-week open label phase, 103 did not complete
it (including 7 deaths = 0.6%, 17 with worsening CHF = 1.4%). Thus, 696
patients were randomized to carvedilol (mean dose 45 mg), 398 to placebo.
Median duration of therapy was 6.5 months.
Main characteristics of both groups were similar. For carvedilol, these
NYHA class IV: 3% (the remainder were class II and class III). Average
Ischemia as cause of CHF: 48%
BP at baseline 116/72; HR 84.
Rx: Digoxin 91%; diuretic 95%; ACE inhibitor 95%; other vasodilator 32%
(since the carvedilol group was larger than the placebo group, percentages
are more meaningful here)
Mortality: In the placebo group, there were 7.8% deaths (31), vs.
3.2% in the carvedilol group (22) (p<0.001), which represented a 65%
decrease in the risk of dying (95% CI 39-80%). Most of the deaths were
due to sudden death or progressive CHF. These results held across most
subgroups, although the numbers were not always sufficient to reach statistical
significance. Of interest, the hazard ratio was 0.61 (95% CI 0.25-1.49)
for patients with baseline HR less than 82, and 0.26 (0.12-0.55) for those
with HR > 82.
Combined morbidity and mortality: Hospitalization for cardiovascular
causes was 14.1% in the carvedilol group and 19.6% in the placebo group.
Combined risk of dying or being hospitalized (taking into account that
a patient who died could not subsequently be hospitalized), was 24.6% placebo,
Side-effects and discontinuation: The most common side-effect, more
frequent in the carvedilol group than the placebo group, was dizziness,
which seldom led to discontinuation of the study drug. The study drug was
discontinued for adverse effects in 7.8% of the placebo group and 5.7%
of the carvedilol group. The only reason for discontinuation that seems
to have been more frequent in the carvedilol group was bradycardia (6 vs
Methodologic issues: In an accompanying editorial, Pfeffer and Stevenson
point out that seven deaths occurred during the two-week run-in phase,
when patients were being treated with 6.25 mg of carvedilol in open label
fashion. Had these deaths been attributed to carvedilol, they would have
accounted for 25% of carvedilol deaths and would have greatly diminished
the calculated benefit of the drug. The authors, in their discussion, state
that this 0.6% death rate during the two weeks is not excessive; they attribute
it to the natural history of the disease, and note that the death rate
during the prior three week screening period was 1.7%, in the absence of
The editorial also comments on the relatively small number of deaths
(53) and short follow-up time, compared to trials that looked at the benefits
of ACE inhibitors in CHF.
Finally, both the authors and the editorial point out that these results
only apply to patients with mild and "moderate" CHF, as defined by this
study, since too few patients were in the severe CHF group.
Bottom line: In patients with mild or moderate CHF, 6-month treatment
with carvedilol reduced mortality from 7.8% to 3.2%. This corresponds to
an absolute risk reduction of 4.6% which yields a NNT of 22 (need to treat
22 patients for 6 months to prevent one death).
Whether or not these results will apply to other beta-blockers remains
unclear, and is an important issue, since newer agents such as carvedilol
and bucindolol will be much more expensive than older ones like metoprolol.
Trials are ongoing looking at some of these agents. This is just another
example of the problem posed by trials looking at newer agents of an established
class. The manufacturer of the product can say "only me" with some scientific
justification, while common sense, economics and the manufacturers of rival,
same-class drugs will say "me too". What is needed are more trials looking
at head-on comparisons of drugs within the same class, once efficacy has
been established for one of them.
The patients to whom these results apply, and the exact manner of initiating
therapy remain to be elucidated. Patients with severe CHF were not adequately
represented. The results of the study seem to indicate that higher baseline
heart rate implies greater benefit. Beta-blockers are not yet FDA approved
for the treatment of congestive heart failure. Nevertheless, this is the
first study to clearly indicate a mortality benefit of a beta-blocker in
the treatment of congestive heart failure and represents an important advance.
October 24, 1996
to the Editor about this article from the NEJM website.
Journal Club on the Web main
Submit a comment about this